The Inuits are frequently described as being somehow protected from CVD through their traditional diet and lifestyle(Reference Chateau-Degat, Dewailly and Louchini1). Yet, the recent adoption of a Westernised lifestyle has been associated with an increase in the prevalence of obesity and other diet-related disorders, including type 2 diabetes, hypertension and CVD(Reference Chateau-Degat, Dewailly and Louchini1–Reference Chateau-Degat, Dewailly and Noel3). Blood lipid levels play a causal role in the development of CVD. A meta-analysis has shown that LDL-cholesterol (LDL-C) and HDL-cholesterol (HDL-C) are independently associated with CVD risk(Reference Lewington, Whitlock and Clarke4). Moreover, there is evidence that an increase in blood TAG concentration is also an independent risk factor for CVD(Reference Bansal, Buring and Rifai5). Further, there exists strong evidence to suggest that plasma lipid response to dietary fat content is, to a large extent, genetically controlled(Reference Rudkowska and Vohl6). Thus, common genetic polymorphisms may render an individual more or less responsive to changes in dietary fat intake. Moreover, an individual's response to dietary fat is probably due to a combination of polymorphisms from various genes rather than a single polymorphism(Reference Rudkowska and Vohl6). Overall, gene–diet interactions may provide important insights into the inter-individual variability observed in plasma lipid levels and thus on the risk of CVD.
The genetic background of the Inuit population has also been shown in previous studies to be protective; however, others have reported potentially detrimental effects for CVD susceptibility(Reference Hegele, Young and Connelly7, Reference de Maat, Bladbjerg and Johansen8). The interactions between polymorphisms, dietary fat intake and plasma lipid levels in the Inuit population may thus be different from those previously observed in other populations. Therefore, the main objective of the present study was to examine the associations between thirty-five known polymorphisms (SNP) in twenty candidate genes related to lipid metabolism, the intake of total fat (TotFat) or saturated fat (SatFat), and plasma lipid levels (total cholesterol (TC), LDL-C, HDL-C and TAG) in the Inuit population.
Subjects and methods
Subjects
Data were collected in the framework of the Nunavik Inuit Health Survey ‘Qanuippitaa? How are we?’ conducted among the Inuit of Nunavik in 2004. The survey covered all fourteen Nunavik communities. In the winter of 2009, participants from the Qanuippitaa Nunavik Health Survey in 2004 were revisited for the additional genetic component. The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Ethics Committee on Research of Laval University and the Québec Public Health Ethics review board. Written informed consent was obtained from all subjects.
Anthropometric measurements
Height was measured using a rigid square and a measuring tape in a standardised standing position, with the participant's back against the wall looking straight ahead, with arms hanging down freely. Body weight was measured with a scale (Tanita TBF-300; Tanita Corporation). BMI was calculated as body weight (kg)/height (m2), and internationally recommended cut-offs were used(Reference Kuczmarski and Flegal9).
FFQ
A validated FFQ(Reference Santé-Québec10) was administered to the participants. The FFQ covered sixty-nine food items and beverages. Foods were divided into two major groups. The first group, ‘country foods’, refers to food items derived from fishing, hunting and gathering, recorded for each of the four seasons of the year before the interview. The second group, ‘store-bought foods’, refers to most store-bought foods imported from southern regions and consumed during the month before the survey. Pre-defined serving sizes were included in the questionnaire and a corresponding food model was shown to the respondents. Analysis of the FFQ data provided estimates of consumption frequency and the usual intake in grams of country foods on a daily, weekly, monthly, seasonal or an annual basis. Daily food intakes were calculated on an annual basis by multiplying food consumption frequency by intake in grams for each food.
Plasma blood sampling
Blood samples were collected from an antecubital vein into vacutainer tubes containing EDTA. Plasma was separated by centrifugation (2500 g for 10 min at 4°C) and samples were portioned and frozen for subsequent measurements. Plasma TC and TAG levels were determined using enzymatic methods using a Hitachi 917 autoanalyser and reagents from Roche Diagnostics(Reference McNamara and Schaefer11, Reference Burstein and Samaille12). The HDL-C fraction was obtained directly by selectively inhibiting the reaction with other lipoproteins. LDL-C was calculated with the Friedewald formula(Reference Friedewald, Levy and Fredrickson13). apoB and apoA1 concentrations and lipoprotein fractions were measured in plasma by the rocket immunoelectrophoretic method of Laurell(Reference Laurell14). Lipid analyses were performed at the Centre de Recherche sur les Maladies Lipidiques (Centre de Recherche du CHUQ, Quebec, QC, Canada).
DNA extraction and genotyping
DNA was extracted from 100 μl of buffy coat using the QIAamp 96 DNA Blood Kit (Qiagen, Inc.). The Quant-iT PicoGreen® dsDNA Assay Kit (Invitrogen) was used to quantify DNA. DNA was analysed with TaqMan® Pre-Designed SNP Genotyping Assays according to the manufacturer's instructions at the McGill University/Génome Québec Innovation Center (Montreal, Canada).
From a literature review, thirty-five polymorphisms in twenty genes relating to lipid metabolism were identified, namely: angiotensin I-converting enzyme (ACE) I/D alleles (rs4343), rs4341; angiotensinogen (AGT) M235T (rs699), T174M (rs4762), A-20C (rs5050), A-6G (rs5051); apoA-I (APOA1) G-75A (rs670), 84T/C (rs5070); apoA-IV (APOA4) Asn147Ser (rs5104); apoA-V (APOA5) T-1131C (rs662799), -3A/G (rs651821), S19W (rs3135506), Gly185Cys (rs2075291); apoB (APOB) XbaI (rs693); apoC-III (APOC3) 3238C>G (rs5128), T-455C (rs2854116); apoE (APOE) Cys112Arg (rs429358), Arg158Cys (rs7412), G-219T (rs405509); ATP-binding cassette, subfamily A (ABC1), member 1 (ABC1A) Arg219Lys (rs2230806); cholesteryl ester transfer protein (CETP) TaqIB (rs708272), C-629A (rs1800775), C-4502T (rs183130), G-971A (rs4783961), Ile405Val (rs5882); cytochrome P450, family 1, subfamily A, polypeptide 1 (CYP1A1) Msp1 (rs4646903), A4889G (rs1048943); fat mass and obesity associated (FTO) rs9939609; glucokinase (hexokinase 4) regulator (GCKR) rs780094; insulin-induced gene 2 (INSIG2) rs7566605; lipoprotein lipase (LPL) HindIII (rs320), Ser447Ter (rs328); hepatic lipase (HL or LIPC) C-514T (rs1800588), G-250A (rs2070895); methylenetetrahydrofolate reductase (NAD(P)H) (MTHFR) C677T (rs1801133); paraoxonase 1 (PON1) L55M (rs854560), Gln192Arg (rs662), C-107T (rs705379); PPPARγ2 (PPARG2) Pro12Ala (rs1801282), − 681C/G (rs10865710); transcription factor 7-like 2 (TCF7L2) C47833T (rs7903146).
Statistical analysis
Hardy–Weinberg equilibrium was tested with the Allele Procedure in SAS, version 9.2 (SAS Institute, Inc.). Distribution of alleles was compared between the Inuit and Caucasian populations using the CEU data from the National Center for Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov/) and analysed using the Fisher exact test.
Data are presented as means with their standard errors. Further, variables were checked for normality of distribution using skewness and the kurtosis values. TAG levels were log-transformed before analyses to normalise their distribution. A regression model was used to evaluate the effect of each polymorphism, the fat intake (either TotFat or SatFat as a continuous variable) and the polymorphism × fat intake interaction effect, adjusted for the effects of age, sex and BMI on each of the lipid variables. Second, a regression model adjusted for the effects of age, sex, BMI as well as energy intake was carried out on each of the lipid variables. Then, the regression β-coefficient (β) was derived to estimate the phenotypic difference imparted by each of the genotype × fat intake interaction effects. In addition, to test for the additive effects of multiple polymorphisms, a ‘risk score’ was calculated based on the number of risk genotypes an individual carried from all the significant polymorphism × fat intake interactions for each lipid parameter. A statistical model was used to evaluate the effect of the ‘risk score’, fat intake (either TotFat or SatFat as a continuous variable) and the risk score × fat intake interaction effect, adjusted for the effects of age, sex and BMI on each of the lipid variables. Again, the regression β was calculated to estimate the phenotypic difference contributed by the number of risk genotypes and fat intake interaction effects. Statistical analyses were performed with SAS statistical software, version 9.2 (SAS Institute, Inc.). Statistical significance was defined as P≤ 0·05.
Results
Subject characteristics
In total, 677 households were contacted, of which 521 agreed to participate including 1056 individuals who signed a consent form and 917 who agreed to the collection of blood samples. In addition, 769 participants completed the FFQ. For the additional genetic component, 658 of the original participants gave consent. Consequently, statistical analyses were done for all individuals where plasma lipid levels, nutritional intake and genotypes were available. Therefore, 553 participants, including 251 men and 302 women, with an average age of 37 (sem 0·8) years and with an average BMI of 28 (sem 0·3) kg/m2 were included. Collectively, the Inuit population had a relatively favourable plasma lipid profile (Table 1) compared with an optimal profile(15). Comprehensive results from blood lipid levels in the whole Inuit population have been published previously(Reference Chateau-Degat, Dewailly and Louchini1). Further, FFQ data demonstrate that the average total dietary fat intake was 76 and 63 g/d (approximately 26 % of energy intake per d) in men and women, respectively, including 27 and 22 g/d of SatFat (approximately 9 % of energy intake per d) in men and women, respectively, for this particular subset of the population.
Table 1 Baseline characteristics of the study subjects (Mean values with their standard errors)
Genotypic characteristics
Genotype frequencies did not deviate from those predicted by the Hardy–Weinberg equilibrium except for the following three SNP: LPL HindIII (rs320), LPL Ser447Ter (rs328) and TCF7L2 C47833T (rs7903146), which were excluded from the statistical analyses. The minor alleles of the SNP AGT T174M (rs4762), AGT A-20C (rs5050), APOA5 Gly185Cys (rs2075291), APOA5 S19W (rs3135506), APOE Cys112Arg (rs429358), FTO (rs9939609) and TCF7L2 C47833T (rs7903146) had low frequencies ( < 0·10) in the Inuit population and were also excluded from the analyses.
In addition, allele frequencies were significantly different (P< 0·05) in the Inuit population compared with the Caucasian population for the following eighteen SNP: ABC1A Arg219Lys (rs2230806); AGT M235T (rs699); APOA1 G-75A (rs670); APOA4 Asn147Ser (rs5104); APOA5 -3A/G (rs651821); APOA5 T-1131C (rs662799); APOB XbaI (rs693); APOC3 3238C>G (rs5128); APOE G-219T (rs405509); APOE Cys112Arg (rs429358); CETP TaqIB (rs708272); CETP C-629A (rs1800775); CYP1A1 A4889G (rs1048943); FTO (rs9939609); MTHFR C677T (rs1801133); PON1 L55M (rs854560); PPARG2 − 681C/G (rs10865710); TCF7L2 C47833T (rs7903146) (see Supplementary material 1, available online).
Gene–diet interaction in plasma lipid levels
A total of fourteen interactions between SNP and TotFat intake were observed. Individuals with the T/T genotype of AGT M235T (rs699), the T/T genotype of CETP Ile405Val (rs5882), the G allele of CYP1A1 A4889G (rs1048943), the C allele of PPARG2 − 681C/G (rs10856710) and the T allele of APOB XbaI (rs693) increased their TC to a greater extent than the other genotype groups with higher TotFat intake (Table 2). Similarly, higher LDL-C levels were associated with a higher dietary TotFat in individuals bearing the T/T genotype of AGT M235T (rs699), the T/T genotype of APOE G-219T (rs405509), the G allele of APOA1 G-75A (rs670), the C allele of APOC3 3238C>G (rs5128), the C allele of PPARG2 − 681C/G (rs10856710) and the T allele of APOB XbaI (rs693) (Table 3). Further, lower HDL-C levels were related to a higher TotFat intake in carriers of the C/C genotype of APOA1 84T/C (rs5070) and the T/T genotype of APOE G-219T (rs405509) (Table 4). Carriers of the T/T genotype of LIPC C-514T (rs1800588) had higher TAG levels with a higher TotFat intake (Table 5). The same SNP × TotFat interactions were seen when energy intake was included in the regression model, in addition to age, sex and BMI (data not shown). In addition, a higher number of these aforementioned risk alleles and higher TotFat intake were associated with significantly higher TC, LDL-C or TAG levels, as well as with lower HDL-C levels (Figs. 1–3).
Table 2 Impact of SNP, dietary fat intake and the interaction SNP×dietary fat intake on total cholesterol levels* (β Coefficients with their standard errors)
a,bβ Coefficients with unlike superscript letters for one SNP were significantly different.
* The model includes SNP, fat intake and the interaction term (SNP × fat intake) with adjustment for the effects of age, sex and BMI.
† The β regression coefficients are derived from absolute values.
‡ P values are calculated with normalised values.
§ Genotype frequency is different between the Inuit and Caucasian populations.
Table 3 Impact of SNP, dietary fat intake and the interaction SNP×dietary fat intake on plasma LDL-cholesterol levels* (β Coefficients with their standard errors)
a,bβ Coefficients with unlike superscript letters for one SNP were significantly different.
* The model includes SNP, fat intake, and the interaction term (SNP × fat intake) with adjustment for the effects of age, sex and BMI.
† The β regression coefficients are derived from absolute values.
‡ P values are calculated with normalised values.
§ Genotype frequency is different between the Inuit and Caucasian populations.
Table 4 Impact of SNP, dietary fat intake and the interaction SNP×dietary fat intake on plasma HDL-cholesterol levels* (β Coefficients with their standard errors)
a,bβ Coefficients with unlike superscript letters for one SNP were significantly different.
* The model includes SNP, fat intake and the interaction term (SNP × fat intake) with adjustment for the effects of age, sex and BMI.
† The β regression coefficients are derived from absolute values.
‡ P values are calculated with normalised values.
§ Genotype frequency is different between the Inuit and Caucasian populations.
Table 5 Impact of SNP, dietary fat intake and the interaction SNP×dietary fat intake on plasma TAG levels* (β Coefficients with their standard errors)
a,bβ Coefficients with unlike superscript letters for one SNP were significantly different.
* The model includes SNP, fat intake and the interaction term (SNP × fat intake) with adjustment for the effects of age, sex and BMI.
† The β regression coefficients are derived from absolute values.
‡ P values are calculated with normalised values.
§ Genotype frequency is different between the Inuit and Caucasian populations.
Fig. 1 Effects of multiple significant SNP (taken from Table 2) as a ‘risk score’ in interaction with total fat on total cholesterol levels. Values are means, with their standard errors represented by vertical bars. a,bMean values with unlike letters for ‘at-risk’ groups were significantly different (P= 0·0002).
Fig. 2 Effects of multiple SNP (taken from Table 3) as a ‘risk score’ in interaction with total fat on LDL-cholesterol levels. Values are means, with their standard errors represented by vertical bars. a,bMean values with unlike letters for ‘at-risk’ groups were significantly different (P= 0·05).
Fig. 3 Effects of multiple SNP (taken from Table 4) as a ‘risk score’ in interaction with total fat on HDL-cholesterol levels. Values are means, with their standard errors represented by vertical bars. a,bMean values with unlike letters for ‘at-risk’ groups were significantly different (P= 0·0014).
Similarly, thirteen interactions between SNP and SatFat intake were observed. The pattern of association with SatFat intake was similar to the pattern obtained with TotFat intake (eleven out of fourteen interactions); however, there were a few differences. The individuals with the T/T genotype of AGT M235T (rs699), the C allele of PPARG2 − 681C/G (rs10856710) and the T allele of APOB XbaI (rs693) had higher TC levels with a higher SatFat intake (Table 2). In addition, APOA1 G-75A (rs670) SNP and SatFat determined the TC levels (Table 2). Similarly (except for one SNP), higher LDL-C levels were associated with a higher SatFat intake for the T/T genotype of AGT M235T (rs699), the G allele of APOA1 G-75A (rs670), the C allele of APOC3 3238C>G (rs5128), the C allele of PPARG2 − 681C/G (rs10865710) and the T allele of APOB XbaI (rs693) (Table 3). Lower HDL-C levels were associated with a higher SatFat intake for individuals with the C/C genotype of APOA1 84T/C (rs5070) and the T/T genotype of APOE G-219T (rs405509) (Table 4). Higher TAG levels were associated with a higher SatFat intake in carriers of the T/T genotype of LIPC C-514T (rs1800588) as well as the C/C genotype of ABCA1 Arg219Lys (rs2230806) (Table 5). Again, the same SNP × SatFat interactions were seen when energy intake was included in the model, in addition to age, sex and BMI (data not shown). Similarly to the SNP × TotFat data, a higher number of these aforementioned risk alleles and higher SatFat intake were associated with significantly higher TC, LDL-C and TAG levels, as well as with lower HDL-C levels (data not shown).
Similar statistical analyses were done with apoB100 (apoB) and apoA1 levels, which yielded the same interaction results as seen for LDL-C and HDL-C levels, respectively (data not shown).
Discussion
The present study confirms that there are genetic differences between the Inuit and Caucasian populations; even so, interactions between polymorphisms, dietary fat intake and plasma lipid levels also exist in the Inuit population. More specifically, seven SNP, namely APOB XbAI (rs693), APOA1 − 75G/A (rs670), APOE 219G/T (rs405509), LIPC 480C/T (rs1800588), AGT M235T (rs699), APOA1 84T/C (rs5070) and PPARG2 − 618C/G (rs10865710), in interaction with TotFat and SatFat intake were associated with plasma levels of at least one of the following risk factors for CVD: TC, HDL-C, LDL-C and TAG.
Participants who had the minor allele of APOB XbAI (rs693) demonstrated higher plasma TC and LDL-C levels when consuming a high TotFat and SatFat diet. Numerous studies have previously demonstrated that carriers of the minor allele had elevated LDL-C levels and a less pronounced plasma lipid response to changes in dietary fat intake(Reference Kathiresan, Melander and Guiducci16–Reference Lopez-Miranda, Ordovas and Ostos22). Thus, individuals who carry the minor allele for rs693 in the APOB gene combined with a higher dietary fat intake may be at risk of hypercholesterolaemia.
Similarly, higher TotFat and SatFat intakes were associated with higher plasma LDL-C levels, as well a higher SatFat intake was associated with higher plasma TC, in heterozygotes for APOA1 − 75G/A (rs670). A previous study showed that subjects carrying the mutated allele had higher TC, LDL-C and TAG levels than homozygotes for the wild-type allele(Reference Mata, Lopez-Miranda and Pocovi23). In addition, a diet rich in PUFA has been shown to induce greater plasma TC and LDL-C decreases in heterozygotes than in wild-type subjects compared with a high-SatFat diet(Reference Mata, Lopez-Miranda and Pocovi23). In contrast, an earlier study showed no differences between genotype groups for any lipid variables; yet, the -75G/A polymorphism appears to have an effect on plasma TC, LDL-C and HDL-C responsiveness to increased PUFA in the diet(Reference Ordovas, Corella and Cupples24). Further, the presence of the A allele has previously been associated with increased promoter activity in vitro (Reference Angotti, Mele and Costanzo25). Overall, studies have shown that the response to dietary fat intake may be modified by the − 75G/A (rs670) mutation.
Moreover, in the present study, participants with the T/T genotype of APOE 219G/T (rs405509) exhibited higher plasma LDL-C and lower plasma HDL-C levels when consuming a high-TotFat and -SatFat diet. Similarly, a study demonstrated that carriers of the T allele had higher LDL-C and apoB plasma levels after the SatFat diet compared with G/G homozygotes(Reference Moreno, Perez-Jimenez and Marin26). In addition, carriers of the T allele had a greater decrease in LDL-C and apoB levels when they modified their diet from a SatFat diet to a carbohydrate diet(Reference Moreno, Perez-Jimenez and Marin26). These results suggest that the 219G/T polymorphism may also partially explain inter-individual differences in plasma lipid response to dietary fat intake.
In the Inuit population, carriers of the minor allele of LIPC C-514T (rs1800588) had higher TAG levels when they consumed a high-TotFat and -SatFat diet. Fan et al. (Reference Fan, Raitakari and Kahonen27) also showed that serum TC, HDL-C and TAG levels increased according to the rs1800588 genotype in the order C/C, C/T and T/T in the Finnish population. Further, studies have suggested that the effects of C-514T polymorphism on HDL-C levels were modified by TotFat and SatFat intake(Reference Ordovas, Corella and Demissie28–Reference Zhang, Lopez-Ridaura and Rimm30). Clearly, there exists a gene–diet interaction between plasma lipid levels, dietary fat intake and LIPC C-514T polymorphism.
Further, higher TotFat and SatFat intakes were related to higher TC and LDL-C in carriers of the minor allele of AGT M235T (rs699). Previously, the M235T polymorphism (rs699) in the AGT gene has been related to an increased risk of CVD(Reference Zafarmand, van der Schouw and Grobbee31) via the presence of hypercholesterolaemia. Overall, these results for AGT (rs699) are supported by previous epidemiological studies(Reference Zafarmand, van der Schouw and Grobbee31); however, this gene–diet interaction should be reconfirmed.
In the Inuit population, two polymorphisms, namely APOA1 84T/C (rs5070) and PPARG2 − 618C/G (rs10865710), had opposite effects than described previously in epidemiological studies, which did not take dietary fat intake into consideration. First, carriers of the minor allele of APOA1 84T/C (rs5070) had lower HDL-C levels with a higher TotFat or SatFat intake. In contrast, other studies have shown that the minor C allele was protective: higher HDL-C and lower TAG levels observed in the Japanese population(Reference Shioji, Mannami and Kokubo32, Reference Yamada, Matsuo and Warita33). Second, participants with the wild-type allele of PPARG2 − 618C/G (rs10865710) had higher TC and LDL-C levels when consuming a higher-TotFat and -SatFat diet. Contrary to the present results, the carrier of the mutated allele of PPARG2 − 618C/G was previously associated with a deteriorated lipid profile in a Caucasian population(Reference Bego, Dujic and Mlinar34). Thus, the addition of dietary fat intake with genetic variations may explain discrepancies in plasma lipids levels. In summary, these results suggest that the interaction with dietary fat may determine the risk allele with these two polymorphisms.
Additionally, four SNP, including APOC3 3238C>G (rs5128), CETP I405V (rs5882), CYP1A1 A4889G (rs1048943) and ABCA1 Arg219Lys (rs2230806), showed an interaction with either TotFat or SatFat intake to modulate one plasma lipid variable. A higher TotFat intake was associated with higher plasma LDL-C levels in individuals who carry the wild-type allele of APOC3 3238C>G (rs5128). Previously, the carrier of the mutated allele of APOC3 3238C>G has been associated with elevated plasma TC and TAG levels(Reference Kraja, Province and Straka35, Reference Russo, Meigs and Cupples36). Lopez-Miranda et al. (Reference Lopez-Miranda, Ordovas and Marin37) also demonstrated a decrease in plasma LDL-C concentration in heterozygous subjects after consumption of a diet high in MUFA compared with an increase in plasma LDL-C concentration in wild-type subjects. Additionally, carriers of the I405V (rs5882) polymorphism in the CETP gene had higher plasma TC levels when consuming a high-TotFat diet. Previously, Darabi et al. (Reference Darabi, Abolfathi and Noori38) showed that subjects carrying the mutated allele had a greater reduction in plasma HDL-C levels without changes in TC compared with subjects with the other genotype after a diet low in PUFA:SatFat ratio(Reference Darabi, Abolfathi and Noori38). Thus, these genetic variations may modulate the effect of dietary fat on blood lipids in the Inuit and other populations; yet, more studies are needed to validate these gene–diet interaction effects.
Furthermore, the CYP1A1 A4889G polymorphism (rs1048943) was associated with higher plasma TC levels with a higher TotFat intake. Studies have shown that CYP1A1 is involved in the bioactivation and detoxification of environmental toxins(Reference Sreeja, Syamala and Hariharan39). The biological significance of this interaction needs to be validated; however, it might be of interest to examine the interaction between polymorphisms, dietary fat intake and environmental toxins in the Inuit population as factors influencing plasma lipid levels.
A higher SatFat intake was related to higher plasma TAG levels in carriers of the C allele of ABCA1 Arg219Lys (rs2230806). Genetic variants of ABCA1 have been associated with altered atherosclerosis progression and fasting lipid concentration; however, results from different studies have been inconsistent(Reference Delgado-Lista, Perez-Martinez and Perez-Jimenez40–Reference Kolovou, Kolovou and Marvaki42). Therefore, the impact of this polymorphism may be altered by gene–diet interactions.
Further, a combination of SNP coupled with dietary fat intake generated additive effects on lipid traits, which increases the risk of CVD. Overall, these risk alleles in interaction with dietary fat intake, including TotFat and SatFat, may significantly influence the risk of CVD, especially when the risk allele is in a greater proportion in the Inuit population compared with the Caucasian population. Thus, these results suggest that the environment, such as dietary fat consumed, interacts with genes to produce phenotypic differences. Similar to previous reports(Reference Liao, Lin and Rundek43), each genetic variant had a small-to-moderate effect; however, the combined effect of all significant variants explained a modest extent of lipid variation.
In conclusion, the findings indicate that genetic variants involved in the modulation of lipid metabolism may play an important role in the modulation of cardiovascular health of the Inuit population. The inclusion of environmental factors, such as TotFat and SatFat, to the genetic association study can help to better predict inter-individual variations in plasma lipid levels. Yet, many other SNP exist in lipid metabolism that have not yet been examined. Further, the present study also supports the notion that the multiple-gene approach might provide a better prediction of lipid variations in the population, but this needs to be investigated in a large-scale study. Overall, the assessment of gene–diet interactions may be a valuable tool to predict the impact of dietary changes on plasma lipid levels in order to reduce the risk of CVD.
Acknowledgements
We express our gratitude to the participants for their collaboration. This study was supported by the Ministère de la santé et des services sociaux du Québec, the Nunavik Regional Board of Health and Social Services, the Network of Centres of Excellence of Canada (ArcticNet), the Nasivvik Centre, the Nasivvik ACADRE Inuit Centre, Northern and Indian Affairs (Northern Contaminant Program) and the Canadian Institutes of Health Research (CIHR). I. R. was supported by a CIHR Bisby Postdoctoral Fellowship Award (200810BFE). Y. G. holds a senior research scholarship from the Fonds de la Recherche en Santé du Québec (FRSQ). M.-C. V. holds a Tier 1 Canada Research Chair in Genomics Applied to Nutrition and Health. The authors' contributions are as follows: I. R. performed the statistical analysis, interpreted the data and wrote the paper; V. B., A. D.-L., B. A., Y. G. and M.-L. C.-D. were responsible for the data collection and the analysis including genotyping, nutrient intakes and lipid parameters in the original study; E. D., R. A. H. and M.-C. V. were responsible for the study design; all authors contributed to the critical revision of the manuscript; I. R. and M.-C. V. had primary responsibility for the final content. All authors read and approved the final manuscript. The authors declare that there are no conflicts of interest.
Supplementary material is available online at http://www.journals.cambridge.org/bjn
